Skip to content

Commit

Permalink
Apply clang format
Browse files Browse the repository at this point in the history
  • Loading branch information
@maddyscientist
maddyscientist committed Oct 26, 2023
1 parent 5a10f57 commit ec509e6
Show file tree
Hide file tree
Showing 2 changed files with 37 additions and 30 deletions.
12 changes: 6 additions & 6 deletions include/kernels/spinor_dilute.cuh
View file
Original file line number Diff line number Diff line change
Expand Up @@ -50,7 +50,7 @@ namespace quda {
dilution_size(v.size()),
src(src),
nParity(src.SiteSubset()),
dims(static_cast<const LatticeField&>(src).X()),
dims(static_cast<const LatticeField &>(src).X()),
dilution_block_dims(dilution_block_dims)
{
for (auto i = 0u; i < v.size(); i++) this->v[i] = V(v[i]);
Expand Down Expand Up @@ -111,12 +111,12 @@ namespace quda {

lat_dim_t block_coords;
for (int i = 0; i < coords.size(); i++) block_coords[i] = coords[i] / arg.dilution_block_dims[i];
int block_idx = ((block_coords[3] * arg.dilution_block_grid[2] + block_coords[2]) * arg.dilution_block_grid[1] + block_coords[1])
* arg.dilution_block_grid[0] + block_coords[0];
int block_idx = ((block_coords[3] * arg.dilution_block_grid[2] + block_coords[2]) * arg.dilution_block_grid[1]
+ block_coords[1])
* arg.dilution_block_grid[0]
+ block_coords[0];

for (int i = 0; i < arg.dilution_size; i++) {
arg.v[i](x_cb, parity) = i == block_idx ? src : vector();
}
for (int i = 0; i < arg.dilution_size; i++) { arg.v[i](x_cb, parity) = i == block_idx ? src : vector(); }
} else {
for (int i = 0; i < Arg::max_dilution_size; i++) { // for these types max = actual size
vector v;
Expand Down
55 changes: 31 additions & 24 deletions lib/spinor_dilute.in.cu
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,10 +3,11 @@
#include <tunable_nd.h>
#include <instantiate.h>

namespace quda {
namespace quda
{

template <typename real, int Ns, int Nc>
class SpinorDilute : TunableKernel2D {
template <typename real, int Ns, int Nc> class SpinorDilute : TunableKernel2D
{
std::vector<ColorSpinorField> &v;
const ColorSpinorField &src;
QudaDilutionType type;
Expand All @@ -15,12 +16,9 @@ namespace quda {
template <QudaDilutionType type> using Arg = SpinorDiluteArg<real, Ns, Nc, type>;

public:
SpinorDilute(const ColorSpinorField &src, std::vector<ColorSpinorField> &v, QudaDilutionType type, const lat_dim_t &local_block) :
TunableKernel2D(src, src.SiteSubset()),
v(v),
src(src),
type(type),
local_block(local_block)
SpinorDilute(const ColorSpinorField &src, std::vector<ColorSpinorField> &v, QudaDilutionType type,
const lat_dim_t &local_block) :
TunableKernel2D(src, src.SiteSubset()), v(v), src(src), type(type), local_block(local_block)
{
switch (type) {
case QUDA_DILUTION_SPIN: strcat(aux, ",spin_dilution"); break;
Expand All @@ -31,7 +29,8 @@ namespace quda {
default: errorQuda("Unsupported dilution type %d", type);
}
if (type != QUDA_DILUTION_BLOCK && v.size() != static_cast<unsigned int>(get_size<Ns, Nc>(type)))
errorQuda("Input container size %lu does not match expected size %d for dilution type", v.size(), get_size<Ns, Nc>(type));
errorQuda("Input container size %lu does not match expected size %d for dilution type", v.size(),
get_size<Ns, Nc>(type));

size_t block_volume = 1;
for (int i = 0; i < src.Ndim(); i++) block_volume *= local_block[i];
Expand All @@ -46,19 +45,23 @@ namespace quda {
for (auto i = 0; i < src.Ndim(); i++) {
if (local_block[i] == 0) errorQuda("Dim %d: Dilution block size = 0", i);
if ((src.X(i) * comm_dim(i)) % local_block[i] != 0)
errorQuda("Dim %d: Invalid dilution block size %d for global lattice dim = %d",
i, local_block[i], src.X(i) * comm_dim(i));
errorQuda("Dim %d: Invalid dilution block size %d for global lattice dim = %d", i, local_block[i],
src.X(i) * comm_dim(i));
}
}

apply(device::get_default_stream());
}

template <QudaDilutionType type>
auto constexpr sequence() { return std::make_index_sequence<get_size<Ns, Nc>(type)>(); }
template <QudaDilutionType type> auto constexpr sequence()
{
return std::make_index_sequence<get_size<Ns, Nc>(type)>();
}

template <QudaDilutionType type>
void apply(TuneParam &tp, const qudaStream_t &stream) { launch<DiluteSpinor>(tp, stream, Arg<type>(v, src, local_block, sequence<type>())); }
template <QudaDilutionType type> void apply(TuneParam &tp, const qudaStream_t &stream)
{
launch<DiluteSpinor>(tp, stream, Arg<type>(v, src, local_block, sequence<type>()));
}

void apply(const qudaStream_t &stream)
{
Expand All @@ -76,31 +79,35 @@ namespace quda {
long long bytes() const { return v.size() * v[0].Bytes() + src.Bytes(); }
};

template <int...> struct IntList { };
template <int...> struct IntList {
};

template <typename real, int Ns, int Nc, int...N>
template <typename real, int Ns, int Nc, int... N>
void spinorDilute(const ColorSpinorField &src, std::vector<ColorSpinorField> &v, QudaDilutionType type,
const lat_dim_t &local_block, IntList<Nc, N...>)
{
if (src.Ncolor() == Nc) {
SpinorDilute<real, Ns, Nc>(src, v, type, local_block);
} else {
if constexpr (sizeof...(N) > 0) spinorDilute<real, Ns>(src, v, type, local_block, IntList<N...>());
else errorQuda("nColor = %d not implemented", src.Ncolor());
if constexpr (sizeof...(N) > 0)
spinorDilute<real, Ns>(src, v, type, local_block, IntList<N...>());
else
errorQuda("nColor = %d not implemented", src.Ncolor());
}
}

template <typename real>
void spinorDilute(const ColorSpinorField &src, std::vector<ColorSpinorField> &v, QudaDilutionType type, const lat_dim_t &local_block)
void spinorDilute(const ColorSpinorField &src, std::vector<ColorSpinorField> &v, QudaDilutionType type,
const lat_dim_t &local_block)
{
checkNative(src);
if (!is_enabled_spin(src.Nspin()))
errorQuda("spinorNoise has not been built for nSpin=%d fields", src.Nspin());
if (!is_enabled_spin(src.Nspin())) errorQuda("spinorNoise has not been built for nSpin=%d fields", src.Nspin());

if (src.Nspin() == 4) {
if constexpr (is_enabled_spin(4)) spinorDilute<real, 4>(src, v, type, local_block, IntList<3>());
} else if (src.Nspin() == 2) {
if constexpr (is_enabled_spin(2)) spinorDilute<real, 2>(src, v, type, local_block, IntList<3, @QUDA_MULTIGRID_NVEC_LIST@>());
if constexpr (is_enabled_spin(2))
spinorDilute<real, 2>(src, v, type, local_block, IntList<3, @QUDA_MULTIGRID_NVEC_LIST @>());
} else if (src.Nspin() == 1) {
if constexpr (is_enabled_spin(1)) spinorDilute<real, 1>(src, v, type, local_block, IntList<3>());
} else {
Expand Down

0 comments on commit ec509e6

Please sign in to comment.